All projects Foundation
Logris Verifier
Independent verification of AI work, grounded in reproducible evidence rather than the worker's report.
Stage: Research & early developmentAccess: No public access yet

Logris Verifier explores how to check the claims of AI workers against the actual state of a system, tests and observed behaviour. It should reveal discrepancies and keep the result of the work separate from permission to act on it. The project is in research and early development.
Focus
- Checking claims against reproducible evidence
- Detecting discrepancies between report and reality
- Separating a result from permission to act
What this project does not claim
- Zero hallucinations
- Proof that any AI is safe
- Universal certification
- Direction
- Foundation
- Stage
- Research & early development
- Access
- No public access yet
- Description reviewed

